T-Test help!! Which tail and which type.

J

Jasminee_03

New Member
#1
#1
I'm currently investigating the effects of a strain on fungus on plant growth.
So I have tested the same species of plant under 7 different conditions.
1 = Control (no fungus)
2, 3, 4 = Fungus A at 3 different concentrations.
5, 6, 7 = Fungus B at 3 different concentrations.

I am going to undertake T-Tests to compare my results.
I just want to make sure I have the tails and type of T Test correct.
I have undertaken a 1 tailed paired T - Test.
(because I want to know if the mean of 1 strain is statistically greater than the other).

Is this correct? I ran an unpaired unequal variance T Test to see if the results are different and it gives me massively different results so I want to make sure I've picked the right test before I start my full write up.
 
Miner

Miner

TS Contributor
#2
#2
You have a lot going on in your design (i.e., different fungus types, different concentrations). A t-test is not really appropriate for several reasons, but the biggest issue is the large number of comparisons. This will increase the familywise error rate to an unacceptable level, resulting in false positive results. A 2-way ANOVA is a better approach. Factor 1 is fungus type with 2 levels. Factor 2 is concentration with 4 levels (level 1 = control).

If results are significant, you can follow up with Dunnett's post hoc test to a control.
 
J

Jasminee_03

New Member
#3
#3
Miner said:
You have a lot going on in your design (i.e., different fungus types, different concentrations). A t-test is not really appropriate for several reasons, but the biggest issue is the large number of comparisons. This will increase the familywise error rate to an unacceptable level, resulting in false positive results. A 2-way ANOVA is a better approach. Factor 1 is fungus type with 2 levels. Factor 2 is concentration with 4 levels (level 1 = control).

If results are significant, you can follow up with Dunnett's post hoc test to a control.
Click to expand...
I originally wanted to do an ANOVA but my supervisor told me he wanted me to compare each concentration against each other and wanted me to do T Tests for each
 
Buckeye

Buckeye

Active Member
#6
#6
I agree with Miner. ANOVA is a better approach. And you can investigate the different 'contrasts' that make up the 2-way model.
 
K

kimmer

New Member
#7
#7
I think you can find so many variant software to solve this question. After the software has gone through all the stages of development, it must undergo several series of "exposures" or controlled experimentation. Exposures are used to gather information on user experience with the software. Through these trials, software developers gain confidence in their ability to develop software that is acceptable to a large percentage of their target market. I recommend https://mlsdev.com/services/web-development . During this phase of testing, software developers are trying to achieve specific design goals by implementing specific test cases into the software. In this stage of software testing, the test cases attempt to identify defect characteristics in the software that can lead to users encountering software defects.
 
Last edited:
You must log in or register to reply here.